前立腺がんの高リスク群に対する信頼性の高い生検フリー診断を可能にするデュアルスレッショルド戦略
Yong Liang1,2, Zhiqiang Chen3, Jingliang Zhang4
1Department of Urology, Centre for Leading Medicine and Advanced Technologies of IHM,The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, Anhui, China.
Background:
Several studies have suggested that biopsy-free radical prostatectomy reduces burdens for selected prostate cancer (PCa) patients pre-screened by clinical indicators and PSMA-PET/CT. However, the absence of pre-operative pathology precludes high-risk PCa (HRPCa) identification, crucial for determining the need for extended pelvic lymph node dissection (ePLND). We evaluated whether combining pre-operative clinical indicators and ¹⁸F-PSMA-1007 PET/CT could non-invasively stratify HRPCa risk.
Methods:
This retrospective diagnostic study included a development cohort of 93 patients from USTC center, as well as two external validation cohorts: one cohort of 83 patients from XYH Hospital and another cohort of 117 patients from XJH Hospital. All enrolled patients met the USTC model value ≥ 0.6 criteria (assessed using a nomogram composed of PSAD and PI-RADS grade) and obtained definitive pathological results within 4 weeks after ¹⁸F-PSMA-1007 PET/CT examination. Patients were stratified as HRPCa (D'Amico criteria: tPSA> 20 ng/mL, GS ≥ 8, or ≥ cT2c) or non-high-risk (NHR). Maximum standardized uptake value (SUVmax) and target-to-background ratios (TBRs) of PSMA-PET/CT were quantified. Diagnostic performance for HRPCa was evaluated using ROC analysis. Multivariate logistic regression analysis identified independent predictors.
Results:
In the development cohort (n = 93), SUVmax (AUC = 0.807) and tPSA (AUC = 0.777) demonstrated good discriminatory performance for identifying HRPCa (both p < 0.05). The dual-threshold criterion (SUVmax> 20.9 and tPSA> 15.09 ng/mL) achieved a consistently high PPV (up to 100% in the development cohort and > 91% in external validation) and specificity, significantly outperforming single parameters (p < 0.001). In two external validation cohorts, similar diagnostic performance was observed.
Conclusion:
A dual-threshold non-invasive strategy is established for predicting HRPCa in biopsy-free radical prostatectomy candidates pre-screened by the USTC model, while its applicability to unscreened or low to intermediate-risk populations remains unproven.
